I got three responses to my question and the second response seems to
have hit the nail on the head. My original post is at the bottom. The
Chris Wheeler had the solution (at least it looks like that so far!),
here is his response:
I don't know if this is applicable, as you didn't mention what OS
revision you are
running, but I thought I'd send it anyways (what can it hurt?).
We ran into a similar problem on our Solaris 2.5.1 machines that had
adapters (V 5.0). Basically, the machine would hang at intermittent
with no error messages. The fix was to install patch 104572-08. Since
the systems have been running strong, and we actually see a marked
Let me know if this is helpful, or if you need further info.
MCI WorldCom, Inc.
UNIX System Administrator
Here are the other two responses, which both looked good too:
Peter Polasek <firstname.lastname@example.org
Default routes, and static routes in general, routes don't normally get
deleted. If adding a static route corrects the problem, then my
guess is that you are running a routing daemon (in.routed or in.rdisc)
and some device is generating bad routes. You may want to save the
following for the good and bad cases:
netstat -nr > netstat.good
ifconfig -a > ifconfig.good
traceroute <some host on another network> > trace.good
netstat -I le0 5 <-- verify the errs columns are low
netstat -I nf0 5 <-- verify the errs columns are low
These results should be compared to the bad to see if the
routes are changing or if there are external errors on each
Hope this helps,
Waqar Hafiz <email@example.com>
One thing you could try is to drop down the mtu on fddi interface.
Ethernet has mtu of 1500 so you could set the same mtu for fddi.
You can do this interactively by
ifconfig nf0 mtu 1500 # assumes nf0 is the fddi interface
or at boot time by putting this in /etc/system
My Original Post:
>>> "D. Stewart McLeod" <firstname.lastname@example.org> 10/20/98 10:55am
E3000, 4 CPUs (248 MHz), 1 GB RAM, 10/100Mb/s Ethernet (at 10Mb/s) and a
FDDI interface (primary interface with FDDI s/w SUNWnfr 5.0).
Over the past four months. this server has been dropping off of the
network. Sometimes the only external access to the box is if you are on
the same sub-net and you are able to logon (hideously slow though).
I can get on the box, restart the default route and if I am lucky we are
back in business.
The last couple of times this has happened, we have had to reboot since
it is not possible to even log on to the console. The other problem is
the frequency of this appears to be increasing. No messages in
/var/adm/messages, appears to happen at "random" times, although mostly
it is late at night or early in the morning (don't all our problems
happen then?!). we are also running HP perfview and when I look back in
the logs/graphs for the times when the network dies, it just dies ... no
lead up traffic or anything.
Anybody have any ideas, possible solutions or hints where to look?
-- # D. Stew McLeod ... working at The Boeing Company .. in Renton, WA. #"A pessimist sees the difficulty in every opportunity; # an optimist sees the opportunity in every difficulty." # - Sir Winston Churchill (1874-1965)
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:51 CDT